Skip to content
This repository has been archived by the owner on Feb 27, 2023. It is now read-only.

[bug] dead loop when downloading files caused by that readMd5 is inconsistent with expectedMd5 #68

Closed
lowzj opened this issue Apr 24, 2018 · 9 comments · Fixed by #582
Closed

[bug] dead loop when downloading files caused by that readMd5 is inconsistent with expectedMd5 #68

lowzj opened this issue Apr 24, 2018 · 9 comments · Fixed by #582
Labels
help wanted kind/bug This is bug report for project

Comments

@lowzj
Copy link
Member

lowzj commented Apr 24, 2018 •
edited

No description provided.

@lowzj lowzj added the bug label Apr 25, 2018
@wu1405
Copy link

wu1405 commented Apr 27, 2018

same problem.
piece range:1090519040-1104150527 error,realMd5:70461da8b94c6ca5d2fda3260c5a8c3b,expectedMd5:a65b502d3a94f3da0d2144cd59e0f626,dstIp:

@lowzj
Copy link
Member Author

lowzj commented Apr 28, 2018

@wu1405 Could you paste some logs about your problem?

@zhhsng
Copy link

zhhsng commented May 4, 2018 •
edited

Same problem. My realMd5 is 70461da8b94c6ca5d2fda3260c5a8c3b too!!! NOT coincidence.
I debugged dfget, and found that in core/fetch.py:PowerClient.run() the piece content is:
piece_cont
(Pdb) ['<html>\r\n<head><title>404 Not Found</title></head>\r\n<body bgcolor="white">\r\n<center><h1>404 Not Found</h1></center>\r\n<hr><center>nginx</center>\r\n</body>\r\n</html>\r\n']

@zhhsng
Copy link

zhhsng commented May 4, 2018

I deployed the supernode as docker. And the docker instance has about 3G disk space as the working directory. When the space is insufficent, the space gc will delete the files in qtdown/.
I mounted a bigger volume to the docker and the problem is solved.

@nonly
Copy link
Contributor

nonly commented May 27, 2018

@lowzj same problem, i deploy more than 10 supernode in VM, disk space is enough; only one supernode has the problem;
when peer download files from the problematic supernode , dead loop will be caused;
i suppose dfget can migrate to next node or download whole files from source when md5 is not matched;
do you have any suggestion ?

@xq2248
Copy link

xq2248 commented Jul 7, 2018

@lowzj same problem,too
I deploy df in three centos7.3 physical machines, the first download file failed

@xq2248
Copy link

xq2248 commented Jul 7, 2018

@lowzj the logs:
[2017-09-07 15:28:20,135] INFO sign:98825-1504769300.121 lineno:140 : task url:http://www.taobao.com
[2017-09-07 15:28:20,135] ERROR sign:98825-1504769300.121 lineno:48 : read meta file fail for tag:checkService
Traceback (most recent call last):
File "/root/df-client/component/metafile.py", line 35, in load
assert len(data_sign) == 40, "data_sign length not equal 40"
AssertionError: data_sign length not equal 40
[2017-09-07 15:28:20,149] INFO sign:98825-1504769300.121 lineno:60 : local http result:test.html-98825-1504769300.121@0.0.1 for path:/check/
and cost:0.004
[2017-09-07 15:28:20,150] INFO sign:98825-1504769300.121 lineno:64 : do register to 192.168.1.150,remainder:['192.168.1.150']
[2017-09-07 15:28:20,754] INFO sign:98825-1504769300.121 lineno:93 : do register result:{u'code': 200, u'data': {u'pieceSize': 4194304, u'fil
eLength': -1, u'taskId': u'5806501cbcc3bb92f0b645918c5a4b15495a63259e3e0363008f97e186509e9f'}} and cost 0.604
[2017-09-07 15:28:20,768] INFO sign:98825-1504769300.121 lineno:110 : pull piece task result:{u'msg': u'client sucCount:0,cdn status:RUNNING,
cdn sucCount:0', u'code': 602} and sleep 1.098 ...
[2017-09-07 15:28:21,923] INFO sign:98825-1504769300.121 lineno:60 : local http result:20971520 for path:/rate/ and cost:0.002
[2017-09-07 15:28:21,940] ERROR sign:98825-1504769300.121 lineno:332 : piece range:0-4194303 error,realMd5:64e2a12134ffafcf72bf8719ed58c59d,e
xpectedMd5:14b2bd47fa82d1f517621d25e3743f3f,dstIp:192.168.1.150,total:169
[2017-09-07 15:28:21,992] ERROR sign:98825-1504769300.121 lineno:332 : piece range:0-4194303 error,realMd5:64e2a12134ffafcf72bf8719ed58c59d,e
xpectedMd5:14b2bd47fa82d1f517621d25e3743f3f,dstIp:192.168.1.150,total:169
[2017-09-07 15:28:22,039] ERROR sign:98825-1504769300.121 lineno:332 : piece range:0-4194303 error,realMd5:64e2a12134ffafcf72bf8719ed58c59d,e
xpectedMd5:14b2bd47fa82d1f517621d25e3743f3f,dstIp:192.168.1.150,total:169

@xq2248
Copy link

xq2248 commented Jul 9, 2018

I find the problem.
According to the official docs, you may not change the nginx user, so all request to supernode will be 403 error. Simplely change the nginx user to root in nginx.conf, or change the rights of the work directory

@xq2248
Copy link

xq2248 commented Jul 9, 2018

some kuy have the 404 problem, check your work directory of nginx 8001 port and the supernode work path

@allencloud allencloud added kind/bug This is bug report for project and removed bug labels Oct 23, 2018
@starnop starnop pinned this issue Jan 12, 2019
@starnop starnop unpinned this issue Jan 24, 2019
This was referenced Jun 4, 2019
Merged
Merged
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Assignees
No one assigned
Labels
help wanted kind/bug This is bug report for project
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

feature: handle client error and repair lowzj/Dragonfly
6 participants
@lowzj @wu1405 @allencloud @zhhsng @nonly @xq2248